Báo cáo khoa học: The influence of cold shock proteins on transcription and translation studied in cell-free model systems pdf

The influence of cold shock proteins on transcription and
translation studied in cell-free model systems
Roland Hofweber
1
, Gudrun Horn
1
, Thomas Langmann
2
, Jochen Balbach
3
, Werner Kremer
1
,
Gerd Schmitz
2
and Hans R. Kalbitzer
1
1 Institut fu
¨
r Biophysik und Physikalische Biochemie, Universita
¨
t Regensburg, Germany
2 Institut fu
¨
r Klinische Chemie und Laboratoriumsmedizin, Klinikum der Universita
¨
t Regensburg, Germany
3 Laboratorium fu
¨
r Biochemie, Universita
¨

t Bayreuth, Germany
Organisms have achieved many mechanisms to survive
drastic changes in environmental conditions. Bacteria
are known to respond to alterations like extreme,
unphysiological temperature, pH value, ionic strength
and pressure in a specific regulation of protein synthe-
sis and degradation. In the case of cold adaptation
many bacteria respond to a decrease in temperature
with the reorganization of their protein expression pat-
tern. In general the protein synthesis of bulk proteins
becomes inhibited whereas expression of cold shock
proteins (CSPs) increases rapidly [1].
CSPs have been proposed to act as regulators of
gene expression for specific proteins during the accli-
mation phase after a downshift in temperature [2].
Keywords
cold shock protein; in vitro translation;
RNA chaperone
Correspondence
H. R. Kalbitzer, Universita
¨
t Regensburg,
Institut fu
¨
r Biophysik und Physikalische
Biochemie, Universita
¨
tsstraße 31,
93053 Regensburg, Germany
Fax: +49 941 9432479

Tel: +49 941 9432594
E-mail: hans-robert.kalbitzer@biologie.
uni-regensburg.de
(Received 21 March 2005, revised 7 June
2005, accepted 27 July 2005)
doi:10.1111/j.1742-4658.2005.04885.x
Cold shock proteins (CSPs) form a family of highly conserved bacterial
proteins capable of single-stranded nucleic acid binding. They are suggested
to act as RNA chaperones during cold shock inhibiting the formation of
RNA secondary structures, which are unfavourable for transcription and
translation. To test this commonly accepted theory, isolated CSPs from a
mesophilic, thermophilic and a hyperthermophilic bacterium (Bacillus sub-
tilis, Bacillus caldolyticus and Thermotoga maritima) were studied in an
Escherichia coli based cell free expression system on their capability of
enhancing protein expression by reduction of mRNA secondary structures.
The E. coli based expression of chloramphenicol acetyltransferase and of
H-Ras served as model systems. We observed a concentration-dependent
suppression of transcription and translation by the different CSPs which
makes the considered addition of CSPs for enhancing the protein expres-
sion in in vitro translation systems obsolete. Protein expression was
completely inhibited at CSP concentrations present under cold shock con-
ditions. The CSP concentrations necessary for 50% inhibition were lowest
(140 lm) for the protein of the hyperthermophilic and increased when the
thermophilic (215 lm) or even the mesophilic protein (451 lm) was used.
Isolated in vitro transcription under the influence of CSPs showed that the
transcriptory effect is independent from the rest of the cell. It could be
shown in a control experiment that the inhibition of protein expression can
be removed by addition of hepta-2’-desoxy-thymidylate (dT
7
); a heptanu-

cleotide that competitively binds to CSP. The data are in line with a hypo-
thesis that CSPs act on bulk protein expression not as RNA chaperones
but inhibit their transcription and translation by rather unspecific nucleic
acid binding.
Abbreviations
Bc, Bacillus caldolyticus;Bs,Bacillus subtilis; CAT, chloramphenicol acetyltransferase; CSP, cold shock protein; dT
7
, hepta-2’-desoxy-
thymidylate; IF, initiation factor; Tm, Thermotoga maritima.
FEBS Journal 272 (2005) 4691–4702 ª 2005 FEBS 4691
Many CSPs have been shown to have a high affinity
to single-stranded nucleic acids and therefore are
thought to act on transcription and ⁄ or translation. For
the mesophilic bacterium Escherichia coli it is known
that up to 25 different proteins are newly induced after
shifting the temperature from 37 to 10 °C [3]. CspA is
one of the most prominent CSPs. It has high amino
acid sequence identity to the ‘cold-shock domain’ of
eukaryotic Y-box proteins that are known to bind
DNA and RNA [4]. It is a member of the protein fam-
ily of bacterial CSPs who form a group of highly
homologous 7.4 kDa proteins known to bind to RNA
and ssDNA with distinct sequence specificity [5].
Cold shock induced CSP synthesis itself is controlled
at different levels including transcription, RNA stabil-
ity, translational and post-translational events. The
cspA-mRNA contains an 159 base 5¢-untranslated
region bearing binding sites for regulatory proteins
such as helicases, RNAses, CSPs, initiation factors
(IFs) and others [7]. This region plays also an essential

role in mRNA stability and translation efficiency [8,9].
CSP itself binds to the 5¢-untranslated region of their
own mRNA to destabilize secondary structures of
mRNA therein leading to higher expression yields.
Although the role of CspA in cold adaptation is not
fully understood, it is known that they are up-regula-
ted during the reorganization period (adaptation
phase) following cold shock. Here a transient inhibi-
tion of bulk protein synthesis occurs and the cell
growth stops for several hours. A putative connection
between these stages has not been shown to date [1,6].
The mechanism of this inhibition is not yet known it
could occur at transcriptional and ⁄ or translational
level and is usually assumed to be regulated by specific
interactions with specific factors involved in the regula-
tion of these processes. However, it has also been sug-
gested that inhibition of protein synthesis is the effect
of unspecific binding of CSPs to ssRNA or ssDNA
[5,6]. Using complex cellular systems it is difficult
to distinguish on the molecular level between these dif-
ferent mechanisms of inhibition in an unequivocal
manner.
Cold shock proteins have the putative function as
‘RNA chaperones’ [2,10,11] and are reported to gener-
ally inhibit RNA secondary structure formation that
stabilize RNA by intramolecular base pairings. These
secondary structures, especially hairpin loops, are sup-
posed to be the main reason why many genes can only
be expressed heterologically in suboptimal amounts
[12]. For testing the chaperone hypothesis, for learning

more about the biochemical function of CSPs, and
for potentially improving the performance of the
combined in vitro transcription ⁄ translation system, we
investigated the influence of different CSPs on the cell-
free protein synthesis with E. coli cell free extracts. For
obtaining a more general view of the properties of
CSPs we used CSPs from various organisms, which
differ in their physiological growth temperature:
Thermotoga maritima exhibits optimal growth at
80 °C, Bacillus caldolyticus at 60 °C and Bacillus subtilis
at 35 °C. These representative CSPs were chosen
because they share a high homology and sequence
identity among this protein family.
Using a cell-free expression system it is possible to
control the concentration of CSPs accurately and thus
their concentration dependent effects. In addition,
complex regulatory events, as they occur in complex
cellular systems and can lead to a misinterpretation of
the data, are less likely. Therefore the system can com-
plement data from whole cells by a more quantitative
analysis of the direct action of CSPs on protein synthe-
sis. It is especially well-suited for testing the RNA
chaperone function of the CSPs because RNA secon-
dary structures are particularly critical; the widely used
T7 RNA polymerase processes faster on DNA than
E. coli ribosomes interact with mRNA. Therefore free
mRNA accumulates in these cell-free systems [13] and
excessive secondary structure formation takes place.
Results
Effect of cold shock proteins on the combined

cell-free transcription/translation of proteins
RNA secondary structures, especially hairpin loops,
are supposed to be the main reason why many genes
can only be expressed in suboptimal amounts by com-
bined in vitro transcription ⁄ translation systems. CSPs
have the putative function as ‘RNA chaperones’ [2]
and therefore could possibly enhance the protein
expression.
To test this hypothesis, a gene sequence which
is only weakly expressed in the cell-free system and is
supposed to form very stable secondary structures
is required. In this case a significant increase of the
expression level by the RNA-chaperone activity can be
expected. The human H-Ras protein served here as a
model, as the human ‘wild type’ coding sequence enco-
ded on pET14bRasc¢ (C-terminal truncated H-Ras) or
on pET14bRasfl (full-length Ras) is supposed to form
extensive secondary structures, especially at the
3¢-region of the mRNA. It can be compared with the
chemically synthesized sequence encoded on pK7Ras
which was optimized in terms of expression rate by
silent mutations [14,15]. These mutations were directed
to maximize the cell-free H-Ras expression by reducing
Influence of CSPs on transcription and translation R. Hofweber et al.
4692 FEBS Journal 272 (2005) 4691–4702 ª 2005 FEBS
the number of base-pairs in the RNA structure whilst
influencing other critical factors for protein expression.
The free energy difference between the two most
stable mRNA structures calculated by mfold [16] is
245 kJÆmol

)1
.
Both plasmids, encoding either the wild-type or the
synthetic sequence, were used in a cell-free expression
system, as indicated in the Experimental procedures;
10 lL of the reaction volume were subjected to
SDS ⁄ PAGE. The difference in expression level of
the two versions of H-Ras is shown in Fig. 1. The
synthetic sequence of pK7Ras results in a prominent
protein band on the SDS PAGE whereas the two wild-
type sequences pET14bRasc¢ and pET14bRasfl do not
result in a visible band on the Coomassie stained gel
and can only be detected by western blotting (Fig. 2).
We therefore deployed CSPs to our E. coli cell-free
system to investigate the effects of the putative RNA
chaperones on the protein expression.
CSPs in different concentrations were used to
monitor the concentration dependent effects on gene
expression. We analyzed the influence on expression
rate of H-Ras by western blotting experiments and of
chloramphenicol-acetyltransferase (CAT) by a colori-
metric assay.
The expression of H-Ras from the wild-type coding
sequence is only visible on the western blot
(Fig. 2) and not on a Coomassie-stained SDS ⁄ poly-
acrylamide gel (Fig. 1). With the addition of the CSP
from T. maritima, the expression level of the protein
decreases so that at TmCSP concentrations higher than
200 lm, only traces of H-Ras protein can be detected.
The expression rate of CAT at 37 °C was monitored

under the influence of the CSPs from T. maritima,
B. subtilis and of B. caldolyticus. Having no effect in
low concentrations up to a certain level (see below), the
expression rate declines rapidly by adding more CSP
until no expression of CAT was detectable (Fig. 3).
Fig. 1. In vitro transcription ⁄ translation of different Ras-constructs.
The reaction was performed for 1 h at 37 °C, 10 lL of each reac-
tion mixture were acetone precipitated and subjected to SDS
PAGE. The gel was stained with Coomassie Brilliant Blue. (A)
Molecular mass standard; (B) blank; (C) pK7Ras; (D) pET14bRasc¢;
(E) pET14bRas. The arrow displays the prominent band of H-Ras
expressed from pK7ras.
Fig. 2. Inhibition of Ras expression by TmCSP investigated by
western blot with a-Ras. Cell-free expression of H-Ras encoded on
pET14brasc¢ in 1 h batch reaction under the influence of different
concentrations of TmCSP as indicated. Ten microlitres of each reac-
tion were acetone precipitated and subjected to SDS ⁄ PAGE fol-
lowed by western blotting with a-Ras.
Fig. 3. Concentration-dependent inhibition of CAT- expression by
CSPs from different microorganisms. CAT was expressed under the
influence of CSPs with the indicated concentrations from the differ-
ent organisms. The reaction products were analyzed according to
Shaw [40]. The measured values are displayed as rectangles
(TmCSP), circles (BcCSP) or triangles (BsCSP). The line fit to the
experimental data using Eqn (2) are shown (see Experimental proce-
dures). The obtained apparent dissociation constants K
app
of CSP
(K
app

given in Table 1, the other fit parameters for TmCSP, BcCSP,
and BsCSP, respectively, are c
DNA
total
¼ 335 ± 244 lM, N ¼
4.01 ± 1.35 l
M, K
ribo
¼ 0. 32 ± 1.3 nM,andc
ribo
total
0.69 ± 0.05 lM.
R. Hofweber et al. Influence of CSPs on transcription and translation
FEBS Journal 272 (2005) 4691–4702 ª 2005 FEBS 4693
In detail, the CSP of the hyperthermophilic organ-
ism T. maritima has no significant effect up to a con-
centration of £ 100 lm, then the expression rate
decreases rapidly so that at concentrations ‡ 230 lm
the expression of CAT is fully suppressed. The addi-
tion of the CSP from the thermophilic bacterium
B. caldolyticus shows no significant effect at concentra-
tions £ 200 lm. The expression rate also decreases until
it is not detectable at BcCSP concentrations ‡ 420 lm.
Furthermore, the CSP CspB from the mesophilic bac-
terium B. subtilis shows no effect at concentrations
£ 300 lm followed by an equivalent decrease of protein
expression as described for the other CSPs. After
reaching BsCSP concentrations of ‡ 800 lm no CAT
expression is observable any more. The optimal growth
temperatures of the respective organisms and the

necessary concentrations for 50% inhibition (c
50
)of
gene expression are displayed in Table 1.
A complete quantitative evaluation of the data is
not possible; however, it is worthwhile fitting the data
with plausible models. The simplest model would
assume the interaction with one component of the sys-
tem (a protein, DNA or RNA) which then directly
abolishes the expression. Formally this situation would
be described by Eqn (1) (see Experimental procedures).
It turns out that the initial constant part of the curve
cannot be described sufficiently well with Eqn (1)
whereas the second part of the data can be described
well by this equation. In a more evolved model, CSP
would decrease the transcription by binding to DNA
and the available mRNA would be limiting for the
expression rate. With this model (Eqn 2 in Experimen-
tal procedures) the data are well described and appar-
ent dissociation constants can be determined (Table 1).
Note that the values obtained do not critically depend
on the other free parameters of the model, a property
of the used function. On the other hand, this means
that the other parameters following from the fit of the
data cannot be determined with high accuracy.
Inhibition of transcription by cold shock proteins
To elucidate whether the effect shown above occurs on
the transcriptional or on the translational level, the
system of combined transcription ⁄ translation was
decoupled. The influence on the transcriptional level

was investigated by in vitro transcription of the CAT
gene using T7 RNA polymerase and nucleotides. CAT
was encoded on pK7cat or on pET14bcat when His
6
-
tagged CAT was used for western blotting analysis.
The transcript was isolated by digestion of RQ1
RNAse-free DNAse and was followed by separation of
nucleotides with nucleospin columns. The purified
transcript was analyzed using an AGILENT 2100 Bio-
analyzer. In the absence of CSPs and in the presence
of the equivalent amount of BSA in the according buf-
fer a distinct band of transcription product was also
present as a small amount of undigested plasmid. The
addition of CSP resulted in the loss of these distinct
bands and a distribution of different bands appears
which could not be assigned (data not shown). In a
second test to determine whether transcription is still
working in the presence of CSPs, in vitro transcription
in the presence of [
32
P]CTP[aP] was performed and
subjected to polyacrylamide electrophoresis. The auto-
radiogram is shown in Fig. 4. As a result of this
experiment, no transcription could be observed when
CSPs of T. maritima, B. subtilis or B. caldolyticus were
added in concentrations which led to a complete sup-
pression of protein expression in the combined tran-
scription ⁄ translation assay (Table 1).
Inhibition of translation by CSPs

The effect of CSPs on the translatory process was inves-
tigated by in vitro translation of mRNA encoding chlo-
ramphenicol acetyltransferase. As a starting point, the
set up for the combined transcription ⁄ translation
experiment was used omitting plasmid and T7 RNA
polymerase as the components for transcription. Instead
of these, mRNA was used as template for the transla-
tion. With this system, the effect of the CSPs on the
translational level was studied. In the absence of CSP
the expression rate was approximately 11 lgÆmL
)1
CAT
in 1 h using an mRNA concentration of 33 lgÆmL
)1
.
When CSPs were present in concentrations where
complete inhibition of the combined in vitro
Table 1. Inhibition of CAT-expression by CSPs from different micro-
organisms.
Source
T
opt
(K)
a
c
50
(lM)
b
c
99

(lM)
c
K
app
(lM)
d
K
D
(lM)(dT)
7
K
D
(lM)(dA)
7
T. maritima 353 139.5 230 0.59 0.02
e
8.0
f
B. caldolyticus 333 215.0 420 2.33 –
g
–
g
B. subtilis 308 451.8 800 4.51 0.37
h
–
g
a
In vitro transcription-translation assay was performed at 310 K.
b
Optimal growth temperature.

c
CSP concentration for 50% inhibi-
tion.
d
CSP concentration for virtually complete inhibition (> 99%).
e
Apparent K
D
from the fit of the in vitro data using Eqn (2) (Fig. 3).
The obtained apparent dissociation constants are largely indepen-
dent of the other free fit parameters used on Eqn (2).
f
K
D
for bind-
ing of (dT)
x
or (dA)
7
to TmCSP interpolated to 310 K from the data
of M. Zeeb (Universita
¨
t Bayreuth, Germany; personal communica-
tion) using the relation lnK ¼ –DG
0
⁄ RT.
g
No data available.
h
K

D
for
binding of (dT)
7
to BsCSP interpolated to 310 K from the data of
[17].
Influence of CSPs on transcription and translation R. Hofweber et al.
4694 FEBS Journal 272 (2005) 4691–4702 ª 2005 FEBS
translation ⁄ transcription was observed (Table 1), trans-
lation was completely abolished (Fig. 5).
Reversal of the inhibitory effect of TmCSP
by single-stranded DNA
We suggest that the effects shown above are based on
the binding of the CSPs to DNA and RNA. However,
other possible mechanisms are the inhibition of
transcription and ⁄ or translation by interaction with
proteins involved in this process or an increased degra-
dation of nucleic acids by an increased nuclease activ-
ity. It was shown for TmCSP (M. Zeeb, Universita
¨
t
Bayreuth, Germany; personal communication) that it
binds tightly to the oligodesoxynucleotide hepta-2’-des-
oxy-thymidylate (dT
7
). Therefore, we tested if binding
of this nucleotide to TmCSP can interfere with the sug-
gested DNA or RNA interaction in the combined
transcription ⁄ translation assay. At a TmCSP concen-
tration of 139.5 lm (the TmCSP concentration for

50% inhibition of the protein expression), the protein
expression was tested in dependence on different con-
centrations of dT
7
. The CSP was incubated with the
nucleotide before cell-free expression was performed,
so that part of the TmCSP molecules were inactivated
with regard to their influence on the transcrip-
tion ⁄ translation processes. Within the limits of error
the full expression level could be re-established by add-
ing dT
7
in approximately equimolar concentration to
TmCSP (Fig. 6), indicating that TmCSP is removed
from its binding sites on nucleic acids by dT
7
. At very
high concentrations of the heptanucleotide the expres-
sion level again decreases somewhat, probably because
of unspecific effects of the oligonucleotide with compo-
nents of the transcription ⁄ translation machinery.
Discussion
The expression efficiency of various genes under the
control of a strong promotor is influenced by many
Fig. 5. Influence of CSPs on the in vitro translation of CAT. Activity
level of CAT after in vitro translation of mRNA transcribed from
pET14bCAT in the presence of 8 mgÆmL
)1
BSA (control), 230 lM
TmCSP, 420 lM BsCSP or 800 lM BcCSP. The insert displays a

western blot of these samples. Samples (5 lL) were acetone preci-
pitated and subjected to SDS ⁄ PAGE [41] following immunodetec-
tion of the His
6
-tagged CAT by His-Probe.
Fig. 4. Influence of CSPs on the in vitro transcription of CAT.
In vitro transcription was performed in the presence of
[
32
P]CTP[aP]. Reaction products (equal amounts of radioactivity)
were subjected to denaturing polyacrylamide gel electrophoresis
with following autoradiography (time of exposure: 45 min); lane
A: standard (F·174 DNA ⁄ Hinf I from Promega); lane B: tran-
scription experiment under the influence of buffer A (pH 6.5) with
2mgÆmL
)1
BSA as control protein; lane C: transcription experi-
ment under the influence of buffer B (pH 7.8) with 8 mgÆmL
)1
BSA as control protein; lane D: addition of 230 lM TmCSP in buf-
fer A; lane E: addition of 420 l
M BcCSP in buffer B; lane F:
addition of 800 l
M BsCSP in buffer B.
Fig. 6. Suppression of the CSP inhibition of CAT-expression by
ssDNA. Titration of poly dT7 to the combined transcription ⁄ transla-
tion of pK7CAT in the presence of TmCSP. Poly(dT
7
) of different
concentrations (final concentrations in expression as indicated) was

incubated with TmCSP (final concentration was 139.5 l
M)for
20 min at 25 °C and then subjected to CAT expression.
R. Hofweber et al. Influence of CSPs on transcription and translation
FEBS Journal 272 (2005) 4691–4702 ª 2005 FEBS 4695
different factors. One of the most crucial factors next
to codon usage is the occurrence of intramolecular
RNA base pairings on mRNA encoding the protein of
interest. These secondary structures are able to mask
regulatory sequences on the mRNA, for example the
Shine–Dalgarno sequence [6], so that translation fac-
tors cannot bind easily to mRNA. Another reason for
low expression levels is the increase of energy necessary
for dissolving secondary structures, e.g. hairpin loops
which has to be provided by the translation machinery
[12]. Consequently, minimization of mRNA secondary
structures leads to a higher expression rate. It was
reported that CSPs function as RNA chaperones [2].
These are defined as RNA binding proteins able to
prevent the formation of RNA secondary structures [1]
and therefore highly increasing the accessibility of the
RNA to ribonucleases [2]. On the other hand, it is
reported that the translation activity is increased when
CSPs are present [10,18,19].
More recently it was shown on E. coli that its trans-
lational apparatus undergoes significant modifications
during cold shock, especially concerning the ribosomes
[3]. CspA, the major CSP from E. coli was found to
act as an activating factor after binding to specific
mRNAs [20].

The starting point for our experiments was the idea
to use CSPs from different organisms in cell free
expression systems in order to optimize the efficiency
of our combined transcription ⁄ translation system,
which is based upon E. coli S-30 cell-free extracts. In
these systems mRNA accumulates due to the higher
processivity of the widely used viral T7 RNA poly-
merase in contrast to the slower E. coli ribosomes [13].
This excessive pool of transcripts leads to an enriched
level of RNA secondary structure formation. Thereby
we could check the hypothesis of [6] that CSPs should
act on protein expression at high concentrations in an
inhibiting manner.
As we wanted to study the effect of the addition of
CSPs to our extracts we prepared our extracts from
E. coli BL 21 carefully under conditions where cold
shock conditions did not prevail. As a model system
for characterizing the influence of CSPs on expression,
H-Ras was chosen as it is available in the wild-type
coding sequence and in a synthetic version [14]. This
synthetic version has been optimized via silent muta-
tions in terms of codon usage and minimization of
possible RNA secondary structures. We verified the
difference in the thermodynamics of the RNA secon-
dary structures by predicting the free energy of the
RNA foldings using the mfold program [16]. This pre-
diction results in a thermodynamic stabilization of the
wild-type gene of 245 kJÆmol
)1
with respect to the syn-

thetic gene sequence. The comparison of the energy
dot blots reveals that the wild-type RNA sequence can
form a large number of intramolecular base pairs at
the 5¢-UTR of the sequence, which can be a reason for
reduced translation initiation. In agreement is the fact
that silent mutations in the synthetic sequence result in
a dramatic increase in expression rate compared with
the wild-type sequence, as depicted in Fig. 1.
The effect of cold shock proteins on the
combined transcription and translation
To test the CSPs’ function as RNA chaperones we
used CSPs from T. maritima, from B. caldolyticus and
from B. subtilis. Their respective optimal growth tem-
peratures were 80, 60 and 35 °C (Table 1).
The temperatures for cold shock response of these
organisms differ from 10 to 60 °C (Table 1) and there-
fore different nucleotide binding affinities have to be
expected for these highly homologous CSPs. For com-
parable results we used the standardized temperature
of 37 °C. At this temperature, where the E. coli system
works optimally, three different scenarios are present:
BsCSP experiences physiological conditions, BcCSP
undergoes cold shock conditions (20 °C below optimal
growth temperature) and TmCSP is even below cold
shock conditions (50 °C below optimal growth tem-
perature).
As an easily quantifiable reporter gene assay we used
the expression of CAT in our standard cell-free expres-
sion system. We applied different concentrations of
CSPs to this experiment. Whether a very small increase

in CAT expression does exist for BcCSP and BsCSP,
but not for TmCSP, at low CSP concentrations
(Fig. 3) cannot be decided from our data because it is
clearly not significant with respect to the inherent
experimental errors of our assay. However, the expres-
sion of CAT under the influence of any of the three
used CSPs resulted in a significant, dramatic decrease
of protein synthesis rate (Fig. 3) in contrast to the
expected increase of expression rate by addition of
CSPs following the RNA chaperone theory. The same
inhibitory effect was visible when TmCSP was added
to the cell-free batch expression of H-Ras.
The three different CSPs exhibit different concentra-
tion ranges for the inhibitory effect described above.
The CSP from T. maritima inhibits the expression
process most effectively followed by the CSP from
B. caldolyticus. The protein that needs the highest con-
centration for an efficient inhibition was the CSP from
B. subtilis.
Our observed concentration levels necessary for inhi-
bition of gene expression are in a physiological range
Influence of CSPs on transcription and translation R. Hofweber et al.
4696 FEBS Journal 272 (2005) 4691–4702 ª 2005 FEBS
as confirmed by different studies on E. coli [10,11]. It
was shown that the homologous protein CspA from
E. coli reaches concentrations of up to several per cent
of the total soluble protein during temperature down-
shift. Assuming a total soluble protein concentration
of 200–300 mgÆmL
)1

[21,22], the concentration of
CspA in the cytosol can reach concentrations in the
millimolar range during cold shock. The effects des-
cribed in this work are detectable at concentrations
< 800 lm. Therefore the effects described here do not
only exist in our in vitro system but can also occur in
the original organisms suffering of too low tempera-
tures for optimal growth.
The inhibition of protein expression by CSPs could
be due to an interaction with other proteins or with
nucleic acids. If the observed effect is caused by bind-
ing to nucleic acids it should be influenced by the
competition of suitable oligonucleotides for the CSP
binding sites. For the CSP from T. maritima binding
data studies to ssDNA are available. The oligodesoxy-
ribonucleotide dT
7
showed maximum affinity to
TmCSP with a K
D
value of (4.0 ± 0.2) · 10
)3
lm at
30 °C and a K
D
value of (0.44 ± 0.02) lm at 50 °C
[21]. For smaller oligonucleotides or other homopenta-
nucleotides tested the affinity was substantially smaller.
Our data show that at concentrations which are high
enough for saturating the CSP in the assay the inhibi-

tory effect of the CSPs on transcription and translation
could be reversed (Fig. 6). This means that for the sup-
pression of protein expression an interaction of CSP
with DNA and⁄ or RNA is required and that protein–
protein interaction or increased nuclease activity can
be excluded as the main inhibitory factors for the
expression process. For reversing the inhibition dT
7
concentrations of approximately 150 lm are necessary
analogous to the TmCSP concentration used in this
assay. This implies that the DNA or RNA interaction
sites with CSP in our assay have comparable or lower
affinities for CSP than dT
7
.
The effect of cold shock proteins on transcription
In the isolated system with only polymerase, nucleo-
tides, template DNA and CSPs without cell lysates
and RNases the effect of CSPs on transcription alone
become observable as no other cellular component of
the heterologous expression system is present. In the
absence of CSPs, a prominent band of transcripts was
present together with a continuous distribution of
smaller transcripts resulting of earlier transcription ter-
mination. These shorter transcripts represent only a
very small part of all mRNA generated in this process
as visible on the autoradiogram, especially with regard
to the fact that equal amounts of radioactivity was
applied to each lane of the polyacrylamide gel. When
CSPs are added in concentrations which lead to an

inhibition of the protein synthesis in the combined
expression experiment (Table 1), no RNA is detectable
any more. As no RNase activity is present, a degrada-
tion of mRNA can be neglected, so that we can con-
clude an inhibition of the transcription process. After
incubation of plasmid DNA with CSPs a complete
digestion of the plasmid with the DNase RQ1 cannot
be observed (data not shown) indicating that CSP
binding nearly completely protects the DNA. As CSPs
are known to bind to single-stranded nucleic acids it
seems reasonable to assume that the CSPs bind with
high affinity to single-stranded DNA as present during
transcription and consequently block the transcription
process.
The effect of cold shock proteins on translation
As shown above CSP inhibits transcription in the used
concentration range but it could also interfere with
protein translation. For the investigation of translation
we subjected mRNA encoding for CAT to a S30
lysate. In this system we could analyze the influence of
CSPs on the translation process. The translation of
CAT is clearly inhibited when CSPs are added to the
system. This effect can be interpreted as a binding of
CSP directly to RNA due to the absence of DNA.
This binding property can result in an increased acces-
sibility of ribonucleases to the mRNA or of masking
of regulatory sequences. The described translation inhi-
bition of a non-cold shock mRNA can be combined
with the results of [3]. They found an increase in the
translation of cold shock mRNAs and of cold tolerant

mRNAs with the addition of CspA to a translation
assay. Thus translation of non-CSPs would be sup-
pressed and that of cold shock-related proteins
enhanced.
In their putative function as RNA chaperones, CSPs
should lead to a higher overall protein synthesis rate
as the formation of RNA secondary structures should
be inhibited or melted [23]. This general function des-
cribed by several groups [2,24] cannot be confirmed
with our results. However, the RNA chaperone func-
tion could be limited to a number of genes whose
mRNA shows motifs like cold shock boxes [7] or cold
shock cut boxes [25]. The preferential binding to these
sequence motifs prevents these specific mRNAs from
folding to stable secondary structures so that CSPs act
as RNA chaperones in these proposed cases. Therefore
we can confirm the hypothesis of [6] that high concen-
trations of CSPs lead to an inhibition of general
R. Hofweber et al. Influence of CSPs on transcription and translation
FEBS Journal 272 (2005) 4691–4702 ª 2005 FEBS 4697
protein expression. Here we bring first evidence for this
theory on three different CSPs in a cell-free system
without regulatory events of a living cell and therefore
focussed on direct nucleic acid binding.
In the light of our results the reduced expression of
bulk proteins visible after the high level expression of
CSPs could be a combined effect of the inhibitory
properties on transcription and translation of CSPs at
these high concentrations where even binding motifs of
lower affinity are occupied. The sense of this break-

down in protein synthesis might be to give the cell time
to rearrange the protein expression pattern to the new
environmental conditions.
Apparent dissociation constants for CSP
As we have shown that direct binding of CSPs to
nucleic acid influences the expression and not specific
RNA chaperone activity, we can analyze the binding
properties of the CSPs used in this study. For a quan-
titative evaluation of the in vitro transcription ⁄ transla-
tion curve different models for the description of the
system are plausible. The simplest model assumes a
direct suppression of the steady state protein expres-
sion by binding of CSP to DNA or mRNA when the
mRNA would be the limiting factor. As mentioned, it
already fits well the data except the constant part at
low CSP concentrations. The slopes of the curves in
Fig. 3 (and thus the apparent binding constants) are
different for the three used CSPs. The steepest descent
can be observed with the addition of TmCSP, followed
by BcCSP and BsCSP indicating the highest nucleic
acid binding affinity for Tm CSP followed by BcCSP as
resulting out of the optimal growth temperatures of
their organisms as indicated above. The apparent dis-
sociation constants are also obtained in a more elabor-
ate model which assumes the existence of a limiting
component different to mRNA as it could be for
example the number of translation-active ribosomes.
Only when the available mRNA concentration would
drop below this value would an effect on protein
expression be observed.

The minimum length of DNA or RNA for optimal
binding to BsCSP (and probably for all CSPs) is six to
seven bases. From the tested sequences the heptanucleo-
tide (dT)
7
has the highest affinity to CSP of B. subtilis
[7] and T. maritima. The extrapolated affinities are
about one order of magnitude higher than those
obtained from the fit of the data (Table 1). However,
they are known to drop substantially for shorter stret-
ches of thymine nucleotides and for sequences including
other nucleotides. In pK7CAT the sequence (T)
7
occurs
once, but outside coding sequences. However, in the
Ras gene the largest stretch of poly(T) has a length of
four and occurs twice, in the CAT gene the largest
stretch comprises six nucleotides and occurs once.
These results are in accordance with earlier findings
that during the acclimation phase of cold shock the
protein expression pattern of the organism drastically
changes and that the synthesis of bulk proteins is tem-
porarily dramatically reduced [1]. Our in vitro data
indicate that CSPs are directly involved in this down-
regulation of protein expression by binding to elements
with intermediate specificity which occur in most
genes. According to our data CSP would inhibit the
new transcription of the majority of genes and also
inhibit the translation of still existing RNA as already
postulated [6].

Most probably the affinities are adapted during evo-
lution to their specific temperature ranges as necessary
for the different growth conditions of their source
organisms. At least the affinities of CSP for poly(T) at
a given temperature are positively correlated with the
optimal growth temperature (Table 1). The above
mechanism does not exclude a second mechanism were
the activity of specific regulatory elements in DNA or
RNA leads to an increased expression of some pro-
teins in the cell. The visible high levels of CSPs during
cold shock can result of many different mechanisms,
like differential increase in its mRNA stability [26] and
preferential translation of cold shock mRNAs resulting
out of an increase of the three translation initiation
factors [3,11]. Furthermore it is known for E. coli, that
the 159 nucleotides of the 5¢-UTR of the cspA mRNA
plays a critical role in the cold shock adaptation. In
many mRNAs encoding for CSPs and for cold tolerant
proteins a special sequence motif (the so-called cold
shock box) was found to be responsible for their cold
shock regulation [3,18,19]. Thereby, the 5¢-UTR that is
responsible for the autoregulation of the transient
expression boosts the expression visible in the early
acclimation phase [8,27].
In this particular case, CSPs may function as RNA
chaperones and therefore promote their own expres-
sion. After reaching a distinct concentration level in
the acclimation phase, one could argue that CSPs also
bind to nucleotide sequences of lower specificity.
Therefore, regulatory sequences of other mRNAs are

then silenced. This leads to a general decrease of bulk
protein synthesis as postulated [6] and demonstrated
experimentally in this work.
Concluding remarks
The observed inhibition of protein expression can be
explained in principle on different levels of the
Influence of CSPs on transcription and translation R. Hofweber et al.
4698 FEBS Journal 272 (2005) 4691–4702 ª 2005 FEBS
combined transcription ⁄ translation assay: (a) the
inhibition of transcription; (b) the increased decay of
transcribed RNA; (c) the decreased translation; and
(d) the increased activity of proteolytic enzymes clea-
ving the target protein. Other effects which can occur
in whole cells (e.g. the induction of new proteins inter-
fering with the protein expression) cannot occur in our
isolated in vitro system. The in vitro system also has
the advantage that the different steps of protein
expression can be observed separately. Therefore we
can clearly determine that inhibition occurs on the
level of transcription which can be explained by bind-
ing of the CSPs to single-stranded nucleic acids as
shown by in vitro transcription. As even the digestion
of the plasmid is protected and the inhibitory effect
can be reversed by a competing oligonucleotide we can
exclude decay of the transcribed RNA. In our cell-free
system, we can also observe a decrease of translation
in an isolated in vitro translation assay.
The effects described here cannot be mapped to a
well-defined binding site for CSPs on the plasmid or
on the mRNA as the effects are visible using different

plasmids and different genes lacking defined recogni-
tion sequences in the regulatory regions. The strongest
binding of CSPs to desoxyoligonucleotides has been
determined for dT
7
. As this motif can hardly be found
in any coding sequences (the uncorrected statistical
probability to find it in one gene coding for 100 amino
acids is of the order of 1 ⁄ 50), the CSPs are expected to
bind to more unspecific sequence motifs with lower
affinities. This is consistent with the apparent K
D
values obtained here. At the concentrations of CSP
existing in vivo, the unspecific inhibition of protein
expression observed in the living cell after a cold shock
can be explained by a direct binding of CSP to RNA
and ⁄ or DNA. Besides this, in our fit, TmCSP shows
the tightest binding, followed by BcCSP and then
BsCSP. These relative proportions correspond well to
those expected from their organisms’ growth tempera-
ture and of their fully inhibiting concentrations.
Experimental procedures
Expression and purification of T. maritima CSP
Protein expression of TmCSP was performed as described
[28,29]. The plasmid coding for TmCSP was transformed
into E. coli Rosetta (DE3) pLysS. The cells were grown in
Luria–Bertani medium [30] containing 50 lgÆmL
)1
ampicil-
lin and 68 lgÆmL

)1
chloramphenicol at 37 °CtoD
600
¼ 1.
Protein expression was induced by adding 1 mm isopropyl
thio-b-d-galactoside, and bacterial growth was continued
at 37 °C for 3 h. Purification of the protein was performed
as described previously [28]. To remove the bulk of the
E. coli proteins without significant coprecipitation of
TmCSP, the supernatant was diluted fivefold and heated
to 80 °C for 30 min. Pure TmCSP was obtained after
hydrophobic interaction chromatography at pH 8.0 and
size exclusion chromatography with Superdex 75 (Amer-
sham, Freiburg, Germany). The total yield of TmCSP was
about 15 mgÆL
)1
cell culture. The purified protein was
concentrated by ultrafiltration to a final concentration of
3.3 mm. The used molar extinction coefficient of TmCSP
was 12660 m
)1
Æcm
)1
.
Expression and purification of B. subtilis CSP
A gene encoding BsCSP B was overexpressed using the T7
RNA polymerase promotor system as described [31]. The
plasmid containing the gene for BsCSP was transformed
into E. coli BL21 (DE3) pLysS. The cells were grown at
37 °C in dYT medium containing 25 lgÆmL

)1
chloram-
phenicol and 300 lgÆmL
)1
ampicillin to D
600
¼ 0.8. Pro-
tein expression was induced by addition of 2 mm IPTG
and carried out for 4 h at 37 °C. Cells were harvested and
lysed as described [31] and the supernatant was applied to
a DEAE anion exchange column (Amersham) equilibrated
with 50 mm Tris ⁄ HCl pH 7.8 [32]. Bound protein was
eluted with a linear NaCl gradient from 0 to 600 mm.
BsCSP eluted at approximately 100 mm NaCl. Fractions
containing BsCSP were adjusted to 50% (w ⁄ v) ammonium
sulfate, bound to a butyl-sepharose 4 FF column (Amer-
sham) and washed with 50 mm Tris ⁄ HCl, 50% ammonium
sulfate pH 7.6 to remove bound nucleic acids. Elution
with 50 mm Tris ⁄ HCl pH 7.6 yielded > 95% pure BsCSP,
as judged from SDS ⁄ polyacrylamide gels [33]. After size
exclusion chromatography (HiLoad SuperdexTM 75 prep
grade column; Amersham) in 50 mm Tris ⁄ HCl, 100 mm
KCl pH 7.8, fractions free of nucleic acids were concentra-
ted by ultrafiltration to a concentration of 5.9 mm. From
1 L of cell culture, 5 mg of BsCSP could be prepared.
The molar extinction coefficient of BsCSP was
5800 m
)1
Æcm
)1

[32,34].
Expression and purification of B. caldolyticus CSP
BcCSP was overexpressed in E. coli K38 pGP1-2 containing
the plasmid pBluescriptII SK with the coding sequence for
BcCSP. The cells were grown at 30 ° C in dYT medium con-
taining 25 lgÆmL
)1
kanamycin and 300 lgÆmL
)1
ampicillin
to D
600
¼ 0.8. Protein expression was induced by tempera-
ture shift to 42 °C carried out for 4 h. Cells were centri-
fuged and lysed as described [31]. After cell lysis and
centrifugation the cell-free extract was heated to 65 °C for
40 min to precipitate most of the E. coli proteins. All fol-
lowing steps were carried out at 4 °C according to the puri-
fication of BsCSP described above. The purified protein
R. Hofweber et al. Influence of CSPs on transcription and translation
FEBS Journal 272 (2005) 4691–4702 ª 2005 FEBS 4699
was concentrated by ultrafiltration to a concentration of
3.0 mm. The used molar extinction coefficient of BcCSP
was 7300 m
)1
Æcm
)1
[35].
Template DNA for combined transcription/trans-
lation and in vitro transcription

As template DNA the plasmid pK7CAT [36], pET14b
CAT, pKRAS and pET14bRas fl and pET14bRas c¢
were used. The plasmids were purified from E. coli TG1
using the QIAfilter Plasmid Maxi Kit (Qiagen, Hilden,
Germany).
Reaction conditions for the batch system of the
combined in vitro transcription/translation
The E. coli S30 cell extract used for the cell-free protein
synthesis was prepared according to [37] from E. coli strain
BL21 (Amersham) due to the lack of proteases and T7
RNA polymerase. The T7 RNA polymerase was added in
defined amounts and its preparation was performed accord-
ing to [38]. The system for cell-free transcription and trans-
lation was adopted from [39] with minor modifications. The
standard system without CSPs consisted of 58 mm
Hepes ⁄ KOH pH 7.5, 1.7 mm dithiothreitol, 1.2 mm ATP,
0.9 mm each of CTP, GTP and UTP, 81 mm creatine phos-
phate (CP) (Sigma, St Louis, MO, USA), 250 lgÆmL
)1
cre-
atine kinase (CK) (Roche, Indianapolis, IN, USA), 4.0%
PEG 8000, 0.64 mm 3¢,5¢-cyclic AMP, 68 lml(3)-5-formyl-
5,6,7,8-tetrahydrofolic acid, 170 l gÆmL
)1
E. coli tRNA
from MRE 600 (Roche), 203 mm potassium glutamate,
27.7 mm ammonium acetate, 4.0 mm magnesium acetate,
protease inhibitor cocktail 1· Complete (Roche), 0.5 U
anti-RNAse (Ambion, Austin, TX, USA), 1.0 mm tyrosine,
0.3 mm of each of the other 19 amino acids, 33 lgÆmL

)1
of
the respective plasmid DNA, 140 lgÆmL
)1
T7 RNA polym-
erase, 35.1% (v ⁄ v) S30 extract in volume of 26 lL and
4 lL water to a final volume of 30 lL. The reaction mix-
ture was incubated at 37 °C for 1 h at 500 r.p.m. in a
microtiterplate on a rotary shaker. When CSPs were titra-
ted to the combined transcription ⁄ translation experiment,
the 4 lL of water was replaced by CSP of different concen-
trations.
In vitro transcription
The in vitro transcription was performed using the Ribo-
probe kit (Promega, Madison, WI, USA) and the plasmid
pK7CAT. The kit was used for the generation of template
RNA for in vitro transcription assays and for analysis of
the influence of CSPs at the transcriptory level. In the latter
case the reaction volume was 20 lL, where 3 lL of the
reaction volume were either 50 mm NaH
2
PO
4
, 100 mm
NaCl, 1 mm EDTA, pH 6.5 (buffer A) with or without
TmCSP or 50 mm Tris ⁄ HCl pH 7.8, 100 mm KCl (buffer
B) with or without Bs CSP or BcCSP.
In vitro translation
For determining the influence of CSPs on the transcription
apparatus the system for combined transcription ⁄ transla-

tion was used without T7 RNA polymerase and protease
inhibitor cocktail. mRNA from the transcription experi-
ments was used as template instead of the plasmid DNA.
Concentrations of the CSPs necessary for virtually complete
inhibition were used as indicated in Table 1. The reaction
volume was 30 lL.
Assay of the reaction products
The amount of synthesized CAT protein was quantified by
a colorimetric assay as described by [40]. The proteins were
also analyzed by SDS ⁄ PAGE [41] and western blotting
after acetone precipitation. The transcription products were
detected with an AGILENT 2100 Bioanalyzer (Palo Alto,
CA, USA) and polyacrylamide gel electrophoresis under
denaturing conditions after transcription in the presence of
[
32
P]CTP[aP].
Prediction of RNA secondary structures
The calculation of mRNA secondary structure formation
was performed with the program mfold based on the algo-
rithm described by [16].
Fitting of the binding isotherms
For a direct comparison of CSP from different organisms,
apparent dissociation constants were derived for the inhibi-
tion of the protein expression by CSP using two models. In
the most simple model the inhibition is predominantly due
to the binding of CSP to one component of the system (e.g.
a regulatory element of the plasmid) and switches off its
activity responsible for protein expression. The activity A
(protein expression) is then given by

Aðc
total
CSP
Þ¼A
0
À
c
total
DNA
À K
app
À c
total
CSP
þ
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðc
total
DNA
À K
app
À c
total
CSP
Þ
2
þ 4K
app
c
total

DNA
q
Á
ð1Þ
with A
0
c
total
DNA
as the protein expression (mgÆmL
)1
Æs
)1
)in
the absence of CSP, c
total
CSP
, the total concentration of CSP,
c
total
DNA
, the total concentration of the CSP binding sites on
the plasmid and K
app
the apparent dissociation constant. In a
somewhat more complex description the availability of the
ribosomal translation system is incorporated. The protein
expression is assumed to be proportional to the concentra-
tion c
bound

ribo
, of ribosomal complexes bound to the mRNA,
that is
Influence of CSPs on transcription and translation R. Hofweber et al.
4700 FEBS Journal 272 (2005) 4691–4702 ª 2005 FEBS
Aðc
total
CSP
Þ¼A
1
c
bound
ribo
¼ A
1
À
K
ribo
þ Nc
total
RNA
þ c
total
ribo
À
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðK
ribo
þ Nc
total

RNA
þ c
total
ribo
Þ
2
À 4Nc
total
ribo
c
total
RNA
q
Á
ð2Þ
where A
1
is an appropriate normalization constant, K
ribo
the apparent dissociation constant of the transcription com-
plex from the RNA, c
total
RNA
the total steady state concen-
tration of mRNA, c
bound
ribo
the concentration of bound
ribosomes and c
total

ribo
the total concentration of bound
ribosomes. The RNA concentration c
total
RNA
can be calcula-
ted with Eqn (1) interpreting A
0
c
total
DNA
as the RNA con-
centration in the absence of CSP.
Acknowledgements
This work was supported by the Deutsche For-
schungsgemeinschaft. We thank Christian Herrmann
for helpful discussions and Franz Schmid for providing
the clones for BcCSP and BsCSP. We thank S. Yoko-
jama for helping us initially with the cell-free system
and for providing the plasmids pK7CAT and
pK7RAS.
References
1 Thieringer HA, Jones PG & Inouye M (1998) Cold
shock and adaptation. Bioessays 20, 49–57.
2 Jiang W, Hou Y & Inouye M (1997) CspA, the major
cold-shock protein of Escherichia coli, is an RNA cha-
perone. J Biol Chem 272, 196–202.
3 Giuliodori AM, Brandi A, Gualerzi CO & Pon CL
(2004) Preferential translation of cold-shock mRNAs
during cold adaptation. RNA 10, 265–276.

4 Kloks CP, Spronk CA, Lasonder E, Hoffmann A, Vuis-
ter GW, Grzesiek S & Hilbers CW (2002) The solution
structure and DNA-binding properties of the cold-shock
domain of the human Y-box protein YB-1. J Mol Biol
316, 317–326.
5 Phadtare S & Inouye M (1999) Sequence–selective inter-
actions with RNA by CspB, CspC and CspE, members
of the Cspa family Escherichia coli. Mol Microbiol 33,
1004–1014.
6 Ermolenko DN & Makhatadze GI (2002) Bacterial
cold-shock proteins. Cell Mol Life Sci 59, 1902–1913.
7 Lopez MM, Yutani K & Makhatadze GI (2001) Inter-
actions of the cold shock protein CspB from Bacillus
subtilis with single-stranded DNA. J Biol Chem 276,
15511–15518.
8 Jiang W, Fang L & Inouye M (1996) The role of the
5¢-end untranslated region of the mRNAfor CspA, the
major cold-shock protein of Escherichia coli, in cold-
shock adaptation. J Bacteriol 178, 4919–4925.
9 Xia B, Ke H, Jiang W & Inouye M (2002) The cold
box Stem-loop proximal to the 5’-end of the Escherichia
coli cspA gene stabilizes its mRNA at low temperature.
J Biol Chem 277, 6005–6011.
10 Brandi A, Spurio R, Gualerzi CO & Pon CL (1999)
Massive presence of the Escherichia coli ‘major cold-
shock protein’ CspA under non-stress conditions.
EMBO J 18, 1653–1659.
11 Gualerzi CO, Giuliodori AM & Pon CL (2003) Tran-
scriptional and post-transcriptional control of cold-
shock genes. J Mol Biol 331, 527–539.

12 de Smit MH & van Duin J (1990) Control of prokaryo-
tic translational initiation by mRNA secondary struc-
ture. Prog Nucleic Acids Res Mol Biol 38, 1–35.
13 Spirin A (2002) Cell-Free Translation Systems. Springer,
Berlin.
14 Miura K, Inoue Y, Nakamori H, Iwai S, Ohtsuka E,
Ikehara M, Noguchi S & Nishimura S (1986) Synthesis
and expression of a synthetic gene for the activated
human c-Ha-ras protein. Jpn J Cancer Res 77, 45–51.
15 Kigawa T, Muto Y & Yokoyama S (1995) Cell-free
synthesis and amino acid-selective stable isotope label-
ing of proteins for NMR analysis. J Biomol NMR 6,
129–134.
16 Zuker M & Stiegler P (1981) Optimal computer folding
of large RNA sequences using thermodynamics and
auxiliary information. Nucleic Acids Res 9, 133–148.
17 Lopez MM, Yutani K, Makhatadze GI (1999) Inter-
actions of the major cold shock protein of Bacillus subtilis
CspB with single-stranded DNA templates of different
base composition. J Biol Chem 274, 33601–33608.
18 Brandi A, Pietroni P, Gualerzi CO & Pon CL (1996)
Post-transcriptional regulation of CspA expression in
Escherichia coli. Mol Microbiol 19, 231–240.
19 Graumann P, Wendrich TM, Weber MH, Schro
¨
der K
& Marahiel MA (1997) A family of cold shock proteins
in Bacillus subtilis is essential for cellular growth and
for efficient protein synthesis at optimal and low tem-
peratures. Mol Microbiol 25, 741–756.

20 Goldstein J, Pollitt NS & Inouye M (1990) Major cold
shock protein of Escherichia coli. Proc Natl Acad Sci
USA 87, 283–287.
21 Martin J & Hartl FU (1997) The effect of macromole-
cular crowding on chaperonin-mediated protein folding.
Proc Natl Acad Sci USA 94, 1107–1112.
22 Zimmerman SB & Trach SO (1991) Estimation of
macromolecule concentrations and excluded volume
effects for the cytoplasm of Escherichia coli. J Mol Biol
222, 599–620.
23 Phadtare S, Inouye M & Severinov K (2002) The
nucleic acid melting activity of Escherichia coli CspE is
critical for transcription antitermination and cold accli-
mation of cells. J Biol Chem 277, 7239–7245.
24 Phadtare S, Inouye M & Severinov K (2004) The
mechanism of nucleic acid melting by a CspA family
protein. J Mol Biol 337, 147–155.
R. Hofweber et al. Influence of CSPs on transcription and translation
FEBS Journal 272 (2005) 4691–4702 ª 2005 FEBS 4701
25 Neuhaus K, Anastasov N, Kaberdin V, Francis KP,
Miller VL & Scherer S (2003) The AGUAAA motif in
cspA1 ⁄ A2mRNA is important for adaptation of
Yersinia enterocolitica to grow at low temperature. Mol
Microbiol 50, 1629–1645.
26 Goldenberg D, Azar I & Oppenheim AB (1996) Differen-
tial mRNA stability of the cspA gene in the cold-shock
response of Escherichia coli. Mol Microbiol 19, 241–248.
27 Bae W, Jones PG & Inouye M (1997) CspA, the major
cold shock protein of Escherichia coli, negatively regu-
lates its own gene expression. J Bacteriol 179, 7081–7088.

28 Welker C, Bohm G, Schurig H & Jaenicke R (1999)
Cloning, overexpression, purification, and physicochem-
ical characterization of a cold shock protein homolog
from the hyperthermophilic bacterium Thermotoga mari-
tima. Protein Sci 8, 394–403.
29 Kremer W, Schuler B, Harrieder S, Geyer M, Gronwald
W, Welker C, Jaenicke R & Kalbitzer HR (2001) Solu-
tion NMR structure of the cold-shock protein from the
hyperthermophilic bacterium Thermotoga maritima.
Eur J Biochem 268, 2527–2539.
30 Sambrook J, Fritsch EF & Maniatis T (1989)
Molecular Cloning, a Laboratory Manual, 2nd edn.
Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, NY.
31 Schindelin H, Herrler M, Willimsky G, Marahiel MA &
Heinemann U (1992) Overproduction, crystallization,
and preliminary X-ray diffraction studies of the major
cold shock protein from Bacillus subtilis, CspB. Proteins
14, 120–124.
32 Schindler T, Herrler M, Marahiel MA & Schmid FX
(1995) Extremely rapid protein folding in the absence of
intermediates. Nat Struct Biol 2 , 663–673.
33 Schaegger H & von Jagow G (1987) Tricine–sodium
dodecyl sulphate–polyacrylamide gel electrophoresis for
the separation of proteins in the range form 1–100 kDa.
Anal Biochem 166, 368–379.
34 Zeeb M & Balbach J (2003) Single-stranded DNA bind-
ing of the cold-shock protein CspB from Bacillus subti-
lis: NMR mapping and mutational characterization.
Protein Sci 12, 112–123.

35 Perl D, Welker C, Schindler T, Schroder K, Marahiel
MA, Jaenicke R & Schmid FX (1998) Conservation of
rapid two-state folding in mesophilic, thermophilic and
hyperthermophilic cold shock proteins. Nat Struct Biol
5, 229–235.
36 Kim DM, Kigawa T, Choi CY & Yokoyama S (1996)
A highly efficient cell-free protein synthesis system from
Escherichia coli. Eur J Biochem 239, 881–866.
37 Pratt JM (1984) Transcription and translation. A Practi-
cal Approach (Hemes BD & Higgins SJ, eds), pp. 179–
209. IRL Press, Oxford.
38 Zawadzki V & Gross HJ (1991) Rapid and simple puri-
fication of T7 RNA polymerase. Nucleic Acids Res 19,
1948.
39 Kigawa T, Yabuki T, Yoshida Y, Tsutsui M, Ito Y,
Shibata T & Yokoyama S (1999) Cell-free production
and stable-isotope labeling of milligram quantities of
proteins. FEBS Lett 442, 15–19.
40 Shaw WV (1975) Chloramphenicol acetyltransferase
from chloramphenicol-resistant bacteria. Methods Enzy-
mol 43, 737–755.
41 Laemmli UK (1970) Cleavage of structural proteins
during the assembly of the head of bacteriophage T4.
Nature 227, 680–685.
Influence of CSPs on transcription and translation R. Hofweber et al.
4702 FEBS Journal 272 (2005) 4691–4702 ª 2005 FEBS